descriptive_statistics
2024-12-03
Descriptive Statistics of NYC Shooting Data
library(ggplot2)
library(readxl)
library(readr)
library(tidyverse)
library(dplyr)
library(sf)
library(plotly)
library(geojsonio)
library(knitr)
library(tidyr)# Load the dataset
df_descriptive=read_csv("data_final.csv")
data_final <- read_csv("data_final.csv")Plots that are related to Time
Shooting Incidents by Year in NYC
# Grouping the data by year to count the number of incidents
df_descriptive$Year <- as.factor(df_descriptive$Year)
year_counts <- as.data.frame(table(df_descriptive$Year))
colnames(year_counts) <- c('Year', 'Count')
# Plotting the line chart for shooting incidents by year with data points labeled
ggplot(year_counts, aes(x = Year, y = Count, group = 1)) +
geom_line(color = 'blue') +
geom_point(size = 3) +
geom_text(aes(label = Count), vjust = -0.5, size = 3) +
labs(title = 'Shooting Incidents by Year in NYC',
x = 'Year',
y = 'Number of Shooting Incidents') +
theme_minimal()
The chart shows a stable trend in shooting incidents from 2017 to 2019, followed by a sharp increase in 2020, likely linked to socio-economic factors like the COVID-19 pandemic. Incidents peaked in 2021 and then declined through 2023.
Bar Chart of Shooting Incidents by Month in NYC
# Grouping the data by month
df_descriptive$Month <- as.factor(df_descriptive$Month)
month_counts <- as.data.frame(table(df_descriptive$Month))
colnames(month_counts) <- c('Month', 'Count')
# Plotting the bar chart for shooting incidents by month
ggplot(month_counts, aes(x = Month, y = Count)) +
geom_bar(stat = 'identity', fill = 'skyblue') +
geom_text(aes(label = Count), vjust = -0.5) +
labs(title = 'Bar Chart of Shooting Incidents by Month in NYC',
x = 'Month',
y = 'Number of Shooting Incidents') +
theme_minimal()
The bar chart shows a seasonal trend in shooting incidents in NYC, with a significant increase during the warmer months, peaking in July (1248 incidents). The highest numbers are seen in June, July, and August, while the lowest are in February (494 incidents) and January (617 incidents), suggesting fewer incidents during colder months. The gradual rise from January to July, followed by a decline towards December, indicates a correlation between warmer weather and increased incidents, potentially due to increased outdoor activities.
Density Plot of Shooting Incident Occurrence Time in NYC
# Convert OCCUR_TIME to a proper time format
df_descriptive$OCCUR_TIME <- format(strptime(df_descriptive$OCCUR_TIME, format = "%H:%M:%S"), "%H:%M:%S")
# Plotting the density of the occurrence time
ggplot(df_descriptive, aes(x = as.POSIXct(OCCUR_TIME, format = "%H:%M:%S"))) +
geom_density(fill = 'skyblue', alpha = 0.6) +
labs(title = 'Density Plot of Shooting Incident Occurrence Time in NYC',
x = 'Time of Occurrence',
y = 'Density') +
scale_x_datetime(date_labels = "%H:%M") +
theme_minimal()
The density plot shows changes in shooting incident trends throughout the day in NYC, with two distinct peaks. The first peak occurs in the late night to early morning hours, and the second peak appears in the evening. The lowest density is observed in the late morning to early afternoon.
Comparison of Shooting Incidents During Daytime and Nighttime
sky_dark_counts <- df_descriptive %>%
group_by(Sky_Is_Dark) %>%
summarise(total_incidents = n())
ggplot(sky_dark_counts, aes(x = Sky_Is_Dark, y = total_incidents, fill = Sky_Is_Dark)) +
geom_bar(stat = 'identity') +
scale_x_discrete(labels = c("FALSE" = "Daytime", "TRUE" = "Nighttime")) + # Update x-axis labels
scale_fill_manual(values = c("FALSE" = "skyblue", "TRUE" = "darkblue"),
labels = c("FALSE" = "Daytime", "TRUE" = "Nighttime")) +
labs(title = 'Number of Shooting Incidents by time (Daytime vs. Nigttime)',
x = 'Time of a day(Daytime vs. Nigttime)',
y = 'Number of Shooting Incidents') +
theme_minimal()
The number of shooting incidents when the sky was dark is significantly higher than when it was bright. This might indicate that shootings are more likely to occur during nighttime or low visibility conditions. The higher number of incidents during dark conditions could be due to factors such as reduced visibility, higher activity at night, or fewer people around, making it easier for incidents to occur undetected.
EDA sorted by neighborhood
Table showing Top 10 NTAs with the Highest Total Shooting Incidents Number (2017-2023)
# Table showing Top 10 NTAs with the Highest Total Shooting Incidents
top_10_nta_overall <- df_descriptive %>%
group_by(BORO, NTA) %>%
summarise(total_incidents = n(),
total_population = first(Total_population_nta),
percent_poverty = first(Percent_poverty),
percent_education = first(Percent_education)) %>%
arrange(desc(total_incidents)) %>%
ungroup() %>%
slice(1:10) %>%
select(BORO, NTA, total_incidents, total_population, percent_poverty, percent_education)
# Display the table in the desired format
kable(top_10_nta_overall, digits = 2, caption = "Top 10 NTAs with the Highest Total Shooting Incidents (2017-2023)")| BORO | NTA | total_incidents | total_population | percent_poverty | percent_education |
|---|---|---|---|---|---|
| BROOKLYN | Brownsville | 350 | 60470 | 37.2 | 76.9 |
| BROOKLYN | Bedford-Stuyvesant (East) | 323 | 84653 | 22.2 | 87.8 |
| MANHATTAN | Harlem (North) | 253 | 83327 | 25.1 | 82.9 |
| BROOKLYN | Crown Heights (North) | 235 | 85275 | 20.5 | 84.1 |
| MANHATTAN | East Harlem (North) | 229 | 64655 | 32.3 | 75.9 |
| BRONX | Mott Haven-Port Morris | 207 | 57718 | 43.7 | 62.0 |
| BROOKLYN | Bedford-Stuyvesant (West) | 207 | 89189 | 26.0 | 84.8 |
| BROOKLYN | East New York-New Lots | 205 | 53004 | 25.3 | 82.6 |
| BRONX | Concourse-Concourse Village | 169 | 69387 | 31.9 | 70.5 |
| BRONX | Mount Eden-Claremont (West) | 166 | 49651 | 31.8 | 67.9 |
Table showing Top 10 NTAs with the Highest Annual Incident Rate
# Get the Top 10 NTAs with the Highest Incident Rate by Year
top_10_incident_rate <- df_descriptive %>%
distinct(incident_rate_by_year_nta, .keep_all = TRUE) %>%
filter(NTAType == 'Residential') %>%
arrange(desc(incident_rate_by_year_nta)) %>%
slice(1:10)
# Pull information for the Top 10 NTAs
top_10_nta_incident_rate <- top_10_incident_rate %>%
select(NTA, BORO,Year,incident_rate = incident_rate_by_year_nta,
total_population = Total_population_nta,
percent_poverty = Percent_poverty,
percent_education = Percent_education)
# Display the table in the desired format
kable(top_10_nta_incident_rate, digits = 5, caption = "Top 10 NTAs with the Highest Incident Rate in Each Borough")| NTA | BORO | Year | incident_rate | total_population | percent_poverty | percent_education |
|---|---|---|---|---|---|---|
| Claremont Village-Claremont (East) | BRONX | 2021 | 0.18735 | 24553 | 43.2 | 61.7 |
| Brownsville | BROOKLYN | 2020 | 0.15380 | 60470 | 37.2 | 76.9 |
| Tremont | BRONX | 2020 | 0.11820 | 32150 | 36.3 | 71.4 |
| Inwood | MANHATTAN | 2021 | 0.11571 | 36299 | 15.3 | 77.0 |
| East New York-New Lots | BROOKLYN | 2020 | 0.10754 | 53004 | 25.3 | 82.6 |
| Hunts Point | BRONX | 2021 | 0.10574 | 15131 | 34.3 | 72.2 |
| Mount Eden-Claremont (West) | BRONX | 2021 | 0.10070 | 49651 | 31.8 | 67.9 |
| Brownsville | BROOKLYN | 2022 | 0.09922 | 60470 | 37.2 | 76.9 |
| Brownsville | BROOKLYN | 2021 | 0.09757 | 60470 | 37.2 | 76.9 |
| Tremont | BRONX | 2023 | 0.09642 | 32150 | 36.3 | 71.4 |
Table showing Top 10 total incident NTA (2017-2023) in Each Borough
# Table showing Top 10 total incident NTA in Each Borough
incident_by_nta_borough <- df_descriptive %>%
drop_na() %>%
group_by(BORO, NTA) %>%
summarise(total_incidents = n()) %>%
arrange(BORO, desc(total_incidents)) %>%
group_by(BORO) %>%
slice_max(n = 10, order_by = total_incidents) %>%
select(-total_incidents) %>%
group_by(BORO) %>%
mutate(row_num = row_number()) %>%
pivot_wider(names_from = BORO, values_from = NTA) %>%
unnest(cols = c(BRONX, BROOKLYN, MANHATTAN, QUEENS, `STATEN ISLAND`)) %>%
select(row_num, everything())%>%
slice(1:10)
# Display the table in the desired format
kable(incident_by_nta_borough, caption = "Top 10 NTAs with Total Shooting Incidents in Each Borough (2017-2023)")| row_num | BRONX | BROOKLYN | MANHATTAN | QUEENS | STATEN ISLAND |
|---|---|---|---|---|---|
| 1 | Mott Haven-Port Morris | Brownsville | East Harlem (North) | Far Rockaway-Bayswater | St. George-New Brighton |
| 2 | Mount Hope | Bedford-Stuyvesant (East) | Harlem (North) | Baisley Park | Tompkinsville-Stapleton-Clifton-Fox Hills |
| 3 | Tremont | Crown Heights (North) | East Harlem (South) | Richmond Hill | West New Brighton-Silver Lake-Grymes Hill |
| 4 | Mount Eden-Claremont (West) | East New York-New Lots | Harlem (South) | Jamaica | Mariner’s Harbor-Arlington-Graniteville |
| 5 | Concourse-Concourse Village | East Flatbush-Remsen Village | Washington Heights (South) | Rockaway Beach-Arverne-Edgemere | Annadale-Huguenot-Prince’s Bay-Woodrow |
| 6 | Williamsbridge-Olinville | East New York (North) | Chelsea-Hudson Yards | St. Albans | Port Richmond |
| 7 | Melrose | Bedford-Stuyvesant (West) | Hamilton Heights-Sugar Hill | South Ozone Park | Rosebank-Shore Acres-Park Hill |
| 8 | Belmont | Canarsie | Inwood | Elmhurst | Grasmere-Arrochar-South Beach-Dongan Hills |
| 9 | Longwood | Coney Island-Sea Gate | Chinatown-Two Bridges | South Jamaica | Great Kills-Eltingville |
| 10 | Fordham Heights | East New York-City Line | Washington Heights (North) | Astoria (East)-Woodside (North) | New Dorp-Midland Beach |
EDA related to Victims
Plot of Description of Male and Female Shooting Victims by Age Group and Race
# Plot: Description of Victims
victim_summary <- df_descriptive %>%
filter(VIC_SEX %in% c("M", "F")) %>% # Filter out unknown
group_by(VIC_SEX, VIC_AGE_GROUP, VIC_RACE) %>%
summarise(total_victims = n()) %>%
drop_na()
# Plotting the description of male victims
ggplot(victim_summary %>% filter(VIC_SEX == "M"), aes(x = VIC_AGE_GROUP, y = total_victims, fill = VIC_RACE)) +
geom_bar(stat = 'identity', position = 'dodge') +
labs(title = "Description of Male Shooting Victims by Age Group and Race",
x = "Victim Age Group",
y = "Number of Victims",
fill = "Victim Race") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
# Plotting the description of female victims
ggplot(victim_summary %>% filter(VIC_SEX == "F"), aes(x = VIC_AGE_GROUP, y = total_victims, fill = VIC_RACE)) +
geom_bar(stat = 'identity', position = 'dodge') +
labs(title = "Description of Female Shooting Victims by Age Group and Race",
x = "Victim Age Group",
y = "Number of Victims",
fill = "Victim Race") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
The bar charts show the distribution of male and female shooting victims by age group and race in NYC. In both charts, Black victims are the most affected across all age groups, with the highest number of victims in the 18-24 and 25-44 age groups. For males, the 25-44 age group shows a notable peak, while for females, the same age group also has the highest numbers. White Hispanic victims also show considerable numbers, particularly in the 25-44 age group.
Plot: Bar chart showing all victim’s age
# Plot: Bar chart showing victim's age
ggplot(df_descriptive %>% filter(!is.na(VIC_AGE_GROUP)& VIC_AGE_GROUP != 1022), aes(x = VIC_AGE_GROUP)) +
geom_bar(fill = "steelblue") +
labs(title = "Bar Chart of Victim's Age",
x = "Victim Age Group",
y = "Count of Victims") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
The 25-44 age group has the highest number of victims, significantly more than any other group. The 18-24 age group also shows a considerable number of victims, while the <18 and 45-64 age groups have notably fewer victims. The 65+ age group has the least number of victims.
EDA sorted by NYC BOROs
Process of data manipulation
Plot of Total Number of Incidents Across NYC BOROs (2017-2023)
data_final <- read_csv("data_final.csv")
boro_map_data <- boro_map_data %>%
mutate(
hover_text = paste("Borough:", boro_name, "<br>Total Incidents:", Number_of_Incidents)
)
# Create the interactive plot with click functionality
plot <- plot_ly(
data = boro_map_data,
type = "scattermapbox",
split = ~boro_name, # Separate polygons by boroughs
color = ~Number_of_Incidents, # Color based on the number of incidents
colors = "viridis", # Use a color scale
text = ~hover_text, # Display hover text
hoverinfo = "text",
marker = list(size = 8, opacity = 0.7)
) %>%
layout(
title = "Total Number of Incidents Across NYC BOROs (2017-2023)",
mapbox = list(
style = "carto-positron", # Base map style
center = list(lon = -74.00, lat = 40.7128), # Center map on NYC
zoom = 9
)
)
# Add click functionality to display the borough name and number of incidents
plot <- plot %>%
event_register("plotly_click") %>%
htmlwidgets::onRender("
function(el, x) {
el.on('plotly_click', function(d) {
var point = d.points[0];
var text = point.text;
alert('You clicked on: ' + text);
});
}
")
# Display the interactive plot
plot